OK. The -metan- people are free to make their own
meta-comment, but I think the long-term aim is, more
or less, to unify. Crudely, the original history
is that there were quite independent programming efforts
with some overlap, but some of the original authors
are in effect retired from the fray and evidently happy to
hand on the baton to those working on it now.
Nick
[email protected]
Austin Nichols
> Nick et al.--
> Just to be clear, I am not asking for an evaluation per se, or a
> comparison of strengths and weaknesses, or other matters of opinion,
> but a paragraph stating some obvious differences in implementation,
> and if appropriate, a set of circumstances under which alternative
> programs might be preferred (or produce identical results--a useful
> bit of info for writing a test script).
>
> Perhaps an example [taken from the kernel() option description] from
> the help for -ivreg2- (my go-to reference for good help files) can
> illustrate my idea:
> "
> Stata's official newey implements HAC standard errors based on the
> Bartlett kernel, and requires the user to specify the maximum number
> of lags used and not the bandwidth; see help newey.
> "
> The thing is, there is no official Stata meta-analysis command, and
> there seems to be some unclear degree of overlap in user-written
> commands. I was hoping an author of one (who has posted to the list
> recently) might offer some insight (if not in the help file, perhaps
> on the list).
>
> On 12/12/06, Nick Cox <[email protected]> wrote:
> > I think Austin indirectly raises an interesting
> > although tricky question of principle here. As a program
> > author I am happy to mention other people's programs
> > (a) when I am aware of them and (b) when they
> > are complementary to mine. And I imagine that attitude
> > attracts widespread agreement.
> >
> > But I'm queasier about getting into an
> > evaluation of other people's programs.
> > More prosaically, I just don't see that as a usual
> > expectation what belongs in a help file.
> >
> > It can reasonably be included in a write-up of a program
> > in a normal academic manner, as "fair comment"
> > on the alternatives.
> >
> > Naturally, I don't want to lay down the law about
> > practices and a help file can be used for various
> > useful purposes. I just want to flag that the issue
> > is a little delicate.
> >
> > Nick
> > [email protected]
> >
> > Austin Nichols
> >
> > > Also, I for one would much appreciate a comparison of the various
> > > user-written packages for meta-analysis included in the
> help file, as
> > > a supplement to "Also see" section at the bottom of the help file.
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
